We propose to build a collaborative database of Genome-Wide Association Studies (GWAS) of breast cancer with over 15,000 cases of European ancestry, and appropriate controls, and impute >2.5 million imputed HapmapS SNPs for these cases and controls. We will then rank all genotyped and imputed SNPs based their P values for association with (a) total breast cancer (b) breast cancer at age <50 (primarily based on 3,200 cases and 3,200 controls scanned in the Breast Cancer Family Registry (BCFR)) and (c) women with ER-negative breast cancer (primarily based on 1,600 cases and 1,600 controls scanned in the NCI Breast and Prostate Cancer Cohort Consortium (BPC3) Based on these results we will perform replication studies in two large-scale Consortia (a) cohorts in the BCP3 and (b) case-control studies in the Breast Cancer Association Consortium (BCAC). In the BPC3 we propose to expand the current six component cohorts (13,500 cases and an equal number of controls) by collaborating with the Women's Health Initiative (4,500 cases and 4,500 controls) to increase the BPC3 sample size to 18,000 cases and 18,000 controls. To narrow the interval of linkage disequilibrium observed at each locus we will collaborate with large pooled GWAS of breast cancer in women with African ancestry (n=2,500 cases and 2,500 controls) and Asian ancestry (n=3,500 cases and 3,500 controls). Building on this trans-ethnic collaborative effort and on the resequencing data from the 1,000 Genomes Project, we will choose SNPs for fine mapping of the loci identified in the previous aims, in 10,000 further European ancestry cases and 10,000 controls, and an additional 3,500 Asian cases and 3,500 controls, in order to narrow the number of variants for which evidence of function will be sought in sub-Project 2.